Pascual, Ana M. and Romero, Manuel and Tizani, Walid (2015...

Pascual, Ana M. and Romero, Manuel and Tizani, Walid (2015) Thermal behaviour of blind-bolted connections to hollow and concrete-filled steel tubular columns. Journal of Constructional Steel Research, 107 . pp. 137-149. ISSN 0143-974X

Access from the University of Nottingham repository: http://eprints.nottingham.ac.uk/45075/1/2015_PascualRomeroTizani_ThermalBehaviourOfBlindBoltConnections_JCSR_authorVersion.pdf

Copyright and reuse:

The Nottingham ePrints service makes this work by researchers of the University of Nottingham available open access under the following conditions.

This article is made available under the Creative Commons Attribution Non-commercial No Derivatives licence and may be reused according to the conditions of the licence. For more details see: http://creativecommons.org/licenses/by-nc-nd/2.5/

A note on versions:

The version presented here may differ from the published version or from the version of record. If you wish to cite this item you are advised to consult the publisher’s version. Please see the repository url above for details on accessing the published version and note that access may require a subscription.

For more information, please contact [email protected]

mailto:[email protected]

Pascual A M, Romero M L, Tizani W. Thermal behaviour of blind-bolted connections to hollow and concrete-filled steel tubular columns. Journal of constructional steel research 2015;107:137-149. http://dx.doi.org/10.1016/j.jcsr.2015.01.014

1

Thermal behaviour of blind-bolted connections to hollow and concrete-

filled steel tubular columns

Ana M. Pascual a, Manuel L. Romero a,*, Walid Tizani b

a Instituto de Ciencia y Tecnología del Hormigón (ICITECH), Universitat Politècnica de València, Valencia, Spain

b Department of Civil Engineering, The University of Nottingham,UK * Corresponding author. e-mail address: [email protected]

ABSTRACT

This paper reports on the thermal analysis of blind-bolts connected to concrete filled steel

tube (CFST) and hollow steel section (HSS) columns. The aim is therefore the investigation

of the temperature distribution in the connected sections and the evaluation of the effects due

to concrete filling and anchored bolt extension. For this purpose, experimental and numerical

work was carried out. The test program involved twelve small-scale unloaded specimens

where the variables were: tube section dimensions, type of blind-bolt, and hollow or concrete

filled steel tubes. Results from the experiments revealed the noteworthy effect of concrete on

bolt temperature reduction, the insignificant influence of tube section dimensions, and the

limited impact of embedded bolt extension. Finite element models (FEM) of connections

were developed to simulate the behaviour of tested pieces. Comparison with tests allowed the

calibration of thermal material properties and characteristics of heat flux in interactions.

Furthermore, assessments of heat transfer problem on the simulation of small-scale pieces

extended to the numerical model of the whole endplate connection between an I-beam and a


2

tubular column. Finally, the suitability of simple methods from Eurocode 3 Part 1.2 and other

references to obtain the temperature on the connection was evaluated.

Keywords: Thermal analysis; Blind-bolt; Anchored extensions; Concrete filled steel tube;

Finite element model; Eurocode 3.

NOTATION

A/V Section factor

CFST Concrete filled steel tube

D height of the beam

EC3 Eurocode 3. Part 1.2

EC4 Eurocode 4. Part 1.2

EXP Experimental

EHB Extended Hollo-bolt fastener system

FEA Finite element analysis

FEM Finite element model

fc Compressive cylinder strength of concrete at room temperature

fy Yield strength of structural steel at room temperature

fu Ultimate strength of structural steel at room temperature

HSS Hollow steel section

HB Hollo-bolt fastener system

h depth in the connection

t thickness of steel tube column

UHB Hollo-bolt in an unfilled section

T temperature


3

1. INTRODUCTION

Besides the load bearing capacity, convenient aesthetical appearance and construction

facilities, one of the most valuable advantages of CFST columns is its appropriate behaviour

in fire. The concrete infill acts as a heat sink, which delays the rise of temperatures in the

cross-section, at the same time that steel works as a shield and maintains the integrity of the

element [1]. The thermal analysis of CFST columns has been covered in several researches,

however, their connection with the beam is usually not studied. The lack of knowledge of the

thermal behaviour of connection between the beam and column has made designers solve the

problem by providing the same protection as in the elements joined. Nonetheless, catastrophic

fires [2, 3] have demonstrated the crucial role of connections in the building failure and the

necessity of its further study.

Up to now, the studies of fire behaviour of endplate bolted connections are scarce.

Initially they were limited to joints between I-section steel beams and open sections steel

columns. For instance, Al-Jabri et al. [4, 5] developed experimental and numerical work to

produce relevant data and insight into flush endplate bolted steel connections at elevated

temperatures. On the other hand, Wang et al. [6] and Dai et al.[7] carried out tests and

simulations of 10 medium scale assemblies that included five different types of joints to open

section steel columns. They studied the forces generated in connections during the fire due to

restrained beam deformations.

In connections to CFST or HSS columns, the bolted solution has to use special fasteners

able to be tightened from one side of the tube, called blind-bolts. Several types of blind-bolts

exist, but Hollo-bolt system (Lindapter International, UK) was chosen for this study due to its

easy assembly and its feasibility of resisting bending moments [8]. The Hollo-bolt is mainly


4

made up of three parts (Fig. 1a): a standard bolt, a sleeve with four slots, and a cone with a

threaded hole where the bolt is screwed. The Hollo-bolt works as follows [9]: firstly, the piece

is inserted in the elements to join through clearance holes, then the bolt is tightened, so that

the threaded cone moves against the inner face of the tube expanding the sleeve legs until the

total clamping force is transmitted.

In recent years, the advantages of Hollo-bolt system have raised the interest of

numerous researchers. Elghazouli et al. [10] studied, through experimental tests, angle

connections between beam and tubular unfilled column by using Hollo-bolt under monotonic

and cyclic loads. Their experimental program also included tension tests on the blind-bolts

that gave notice of the tension behaviour of the bolt system. Liu et al [11, 12] examined the

same type of connections first subjected to shear loads and then to axial loads. Column and

angle thickness together with the distance between Hollo-bolt and beam flange, were the

principal parameters that influenced joint capacity and deformation. Their studies comprised

numerical analysis and component characterization with a good adjustment to the

experimental results. In addition, strength and stiffness of Hollo-bolt system in a T-stub

connection was assessed by Wang et al. [13], who aimed to obtain a theoretical expression in

the framework of component method. They noted the higher flexibility introduced by the

sleeve ductile behaviour.

A modification of this type of blind-bolt (Fig. 1b), called the Extended Hollo-bolt, was

developed at the University of Nottingham [8, 14] in order to be used in CFST connections.

The novelty was the use of a longer shank that ended in a screwed nut whose purpose was to

take advantage of the concrete infill to improve the connection stiffness and be able to resist

bending moments. Eight full-scale tests were carried out by Tizani et al. [8] on flush endplate


5

connection to CFST columns with Extended Hollo-bolt, where results proved the

enhancement provided by the bolt anchorage. It avoided stress concentration in the tube and

the sleeve, distributing it within the concrete, at the same time that enabled the full

development of tensile bolt strength. A more detailed study, focused on bolt behaviour was

undertaken by Pitrakkos and Tizani [14] ,who distinguished four mechanisms involved in

Extended Hollo-bolt behaviour: internal bolt elongation, expanding sleeves, shank bond and

anchorage.

Hence, the blind-bolt performance at room temperature has been studied by several

researchers but there is still a gap in its understanding under fire conditions. In this respect,

the European funded project COMPFIRE [15] has provided a helpful insight into the thermal

behaviour of connections to partially encased composite columns and concrete filled tubular

columns. This was preceded by the work of Ding and Wang [16, 17], who tested four

different types of joints to CFST column in fire and one of them used a blind-bolt (the

Molabolt). However, none of the previous researches studied deeply the effect of the blind-

bolt. On the other hand, it is noteworthy the discussion and contribution from Ding and Wang

[17] to simplified temperature calculation methods of Eurocode 3 Part 1.2 (EC3) [18]

This paper deals with the thermal analysis of blind-bolts in endplate connections to HSS

and CFST columns. Small-scale experiments and numerical models to obtain temperature

distribution were conducted. The test program involved twelve specimens with a single blind-

bolt where variables were: section dimensions, Hollo-bolt or Extended Hollo-bolt, and HSS or

CFST column. The aim of the work was to understand and evaluate the effect of these three

variables.


6

Later on, the finite element models (FEM) of the tested specimens were carried out to

simulate the heat transfer of the connections. An analysis of the thermal properties of

materials and their interactions was accomplished by comparison with laboratory data.

Numerical simulations were completed with models of the whole connection beam to column

in order to verify the conclusions from the small-scale models.

Finally, the simplified methods of EC3 [18] and additional proposals developed by three

researches were discussed. The suitability of these methods to calculate the temperature of

blind-bolts in connection to HSS and CFST columns was evaluated by means of comparison

with the experimental and numerical results.

Moreover, this work will provide valuable information for further thermo-mechanical

calculations of these connections.

2. EXPERIMENTAL TESTS OF THE THERMAL RESPONSE

Due to the lack of experimental data about the thermal behaviour of blind-bolts, a test

programme of twelve unloaded specimens (Table 1 and Fig. 2) was undertaken. For the sake

of simplicity, samples with only one fastener were used. As it is shown in Fig. 3, they

consisted of a blind-bolt (Hollo-bolt fastener system) that clamped a steel plate (equivalent to

endplate) and a tubular section (a piece of HSS or CFST column).

2.1. Specimens definition

The experimental program was divided into four series of three specimens each, Fig. 2.

Each series was tested on a different day and included specimens with the same column

section dimensions (Table 1): the first specimen was a connection to a HSS column with a

Hollo-bolt (termed UHB for Unfilled section with Hollo-bolt), the second one was a


7

connection to a CFST column with a Hollo-bolt (termed HB), and in the last example an

Extended Hollo-bolt was used (termed EHB).

Specimens of series 1 and 2 used square column, with dimensions 150x150 mm and

220x220 mm, respectively. On the other hand, series 3 and 4 were rectangular with section

tube of 250x150 mm and 350x150 mm, respectively. The clamped plate had a square section

110 mm long with a thickness of 15 mm for all the examples. The height of the piece of

column was 285 mm, which was previously proved to be large enough not to produce any

interference in the isotherms distribution.

The concrete mixture had normal strength (fc= 30MPa) with calcareous aggregates.

Moisture content was measured by means of cubic specimens (150x150x150 mm), whose

weight was taken before and after being in an oven at 150º C for 28 days. The value obtained

for the moisture content was approximately 7% in concrete weight for all the mixtures.

The steel of tube columns was structural cold formed steel of grade S355, while the

bolts utilised in the fastener system were high strength M16 (diameter of 16mm) grade 8.8,

where the first digit is the value of ultimate strength (fu= 800 MPa) and the second one is the

portion of ultimate strength that indicates the value of yield strength (i.e. fy= 640 MPa). The

shank length varied depending on the type of fastener used, 75 mm for the Hollo-bolt (HB)

and 120 mm for the Extended Hollo-bolt (EHB). Besides, EHB had a standard nut attached at

the end of the bolt shank, Fig 1b.

The standard sleeve was used in all cases. Its specification is provided in the Lindapter

catalogue [9]. Its length depends on the total thickness to fasten, in this case the 41.5 mm long

was used with a strength of 430 MPa.


8

2.2. Test setup and instrumentation

Tests were performed in installations located at AIDICO (Instituto Tecnológico de la

Construcción) in Valencia, Spain. A 5x3 m horizontal furnace was employed, equipped with

16 gas burners that reproduced the standard ISO834 [19] fire curve.

The three specimens of each series were tested at the same time. They were placed on

bases made of heat-resistant concrete with the objective of being at mid-height of the furnace

chamber, as it is shown in Fig. 4. Thick mineral wool was set in both ends of pieces to avoid

the loss of heat through these surfaces.

Type K thermocouples were used to measure the evolution of temperatures at different

points of the bolt and specimen, Fig. 5. Bolt temperature was obtained at least in three

positions: the head of the bolt, the shank next to the tightened cone, and the end of shank

inside the tubular section, indicated with number 1, 2 and 3, respectively. One more

thermocouple was added in Extended Hollo-bolt fasteners, at an intermediate point of the

shank length (EHB7). Positions 1 and 2 were the same for all the bolts, however position 3

varied with the length of the shank, so that, in the Extended Hollo-bolt it was located deeper

inside the tube.

In concrete filled specimens, temperatures were monitored at 11 different locations. Not

only the connection section (Section A-A’) was controlled, but also a section 100 mm from it

(Section B-B’), Fig. 5. The objective was to provide data for a detailed description of the

temperature distribution within the cross section, and also along the piece.

2.3. Experimental results

Comparison between thermocouples measurements provided insight into the heat flow

across the section and the effect of the three variables: tube section dimensions, concrete infill


9

and type of bolt. During the experiments the failure of some thermocouple protections was

detected as an abrupt and unexpected change in the trend of temperature was registered. These

measurements were not taken into account.

Temperature evolution during fire for specimen with HB of series 3 is depicted through

Fig. 6. Thermocouples on the exposed surface reported similar values (HB1 HB4 and HB8)

up to 30 min, after which they started to separate registering a difference not higher than 50ºC

at 50 min. Temperatures in section A-A’ and B-B’ were approximately the same, as it was

proved comparing values HB5 with HB10, HB6 with HB9, and HB7 with HB11. On the other

hand, thermocouples in HB6 and HB9 reached higher temperatures than HB5, HB7, HB10

and HB11, because HB6 and HB9 were closer to the exposed surface. Meanwhile,

temperatures of HB5, HB7, HB10 and HB11 were similar, as they were taken at nearly the

same distance from the exposed surface, i.e. at the same isotherm. Equal field of temperatures

was detected in the rest of specimens of series 1, 2 and 4.

Fig. 7 shows the increment of temperature in three locations of the shank (HB1, HB2,

HB3) for a HB connection to a CFST column for the four tube section dimensions.

Differences not higher than 50ºC were observed between sections. This leads to the

conclusion that temperature distribution in bolts was not affected by the range of dimensions

of the tube sections analysed here. Nevertheless, further test program needs to include a wider

range of section dimensions.

On the other hand, temperatures in connections to HSS were significantly different to

the ones in connections to CFST columns. Fig. 8a and 8b show the temperature in the bolt for

specimens of series 1 and series 3 respectively. In unfilled sections, and after 30 min of fire

exposure, the whole bolt presented approximately the same temperature (UHB1, UHB2,


10

UHB3). Nonetheless, some discrepancy was observed in the values taken for position 1 in

unfilled pieces (HSS) exposed directly to fire. That could be atributted to nonhomogeneus

heat distribution inside the furnace or to problems in thermocouples installation. Conversely,

in connections to CFST, noteworthy differences between points 1, 2 and 3 were observed.

The temperature in the head of the bolt (HB1 and EHB1) after 30 minutes of fire exposure

was 150ºC higher than in position 2 (HB2 and EHB2), and around 400ºC higher with respect

to point 3 (HB3 and EHB3). Moreover, comparing temperatures at the three locations on the

bolts between HSS and CFST connections, it was confirmed that concrete reduced drastically

the temperature. For instance, after 30 min of exposure, bolt location UHB1 in the bolt head

of HSS connection was aproximately 150 ºC hotter than HB1 or EHB1 in CFST connections,

and 250ºC in the case of temperatures at point 2.

The influence of a longer shank embedded in concrete was analysed by comparison of

the two connections to CFST, with Hollo-bolt and Extended Hollo-bolt. No noticeable effects

were detected in the temperatures at point 1 (HB1 and EHB1) and not even at point 2 (HB2

and EHB2). Same patterns were observed in the four sections.

Finally, it was concluded that in unfilled sections (HSS) temperatures increased faster

and reached higher values than in filled sections (CFST). On the other hand, in connections to

CFST, using Hollo-bolts or Extended Hollo-bolts did not produce any significant bolt

temperature alteration except for position 3.

In addition, some specimens were cut after the experiments to identify the bolt

embedment within the concrete and the state of concrete mixture. In Fig. 9, it could be

observed how concrete filled every gap between the parts of the fastener system, not leaving

any empty space. Moreover, the clearance hole to accommodate the sleeve presented also


11

concrete mixture. Coloration in the concrete closer to the tube indicated the chemical changes

produced in the hottest layers of infill.

3. NUMERICAL SIMULATION OF TEST SPECIMENS

3.1. Description of the finite element model

Thermal behaviour of tested pieces was simulated using the finite element analysis

(FEA) package ABAQUS [20]. Three-dimensional numerical models were performed to

represent the heat transfer problem. The three parts of connections were modelled: the hollow

section, the steel plate and the blind-bolt, Fig. 10. Concrete infill was added as a fourth part in

connections to CFST, Fig. 10d.

To model the fastener system, its geometry was simplified, reducing the number of

elements to two: a main element which included shank, bolt head and fastener cone in the

same solid; and the second one represented sleeve with legs in expanded state. The former

changed in connections with Extended Hollo-bolt, where length of shank was larger and

included a screwed nut attached at the end. Fig. 10c shows the parts corresponding with the

blind-bolt of the FEM model. Avoiding complexity, the screw thread in bolt shank was not

considered, nor the hexagonal shape of the bolt head and nut.

Despite the elements reduction, multiple interactions still existed between the several

surfaces of different elements, therefore the FEM required detailed definitions of thermal

contacts and their properties.

Three-dimensional eight-node heat transfer solid elements with thermal degree of

freedom DC3D8 were used. A mesh sensitivity analysis was performed and finer elements


12

were used for bolts (2-5 mm) due to the size of the piece and to guarantee accuracy on the

critical part. The size for the rest of the elements was not higher than 20 mm.

3.1.1. Thermal material properties

Concrete

Temperature-variable thermal properties for concrete, specific heat and thermal

conductivity, were taken from Eurocode 4 Part 1.2 (EC4) [21]. The moisture content is

considered in the peak value of the specific heat at 115ºC. Recommendations give a peak

value of 2020 J/kg K for a moisture content of 3% in concrete weight, and 5600 J/kg K for a

moisture content of 10%. A linear interpolation was used to calculate the peak for

intermediate values. A moisture content of 7% in concrete weight was actually measured for

the tested specimens.

Steel

The temperature dependent thermal properties for structural steel were extracted from

EC3 [18], which are based on mild steel tests. Chemical composition and heat treatment in the

fabrication of high strength steel mean different changes in the internal structure of the

material under elevated temperatures. However, properties of normal strength steel have

traditionally been used to characterise high strength bolts. In that respect, Kodur et al. [22]

proposed new equations for thermal properties of steel bolts of grades A490 and A360.

The effect of considering mild steel properties in contrast with Kodur [22] expressions

for high strength steel was analysed in section 3.2.1 of the present work.


13

3.1.2. Thermal analysis

A nonlinear thermal analysis was performed for each specimen. The thermal load was

transferred to the exposed surfaces by convection and radiation mechanisms. Instead of

applying the standard curve ISO834 [19], the temperature evolution registered in the furnace

during experiments was used, in order not to add more discrepancies in the results, since they

differed slightly.

The parameters that defined the heat transfer problem were adopted from Eurocode 1

Part 1.2 [23]:

� Coefficient of convective heat transfer at the exposed surface: h= 25 W/m2K.

� Configuration factor for radiation at the exposed surface: ϕ=1.

� Stephan-Boltzmann constant: σ=5.67 x 10-8 W/m2K4.

� Emissivity of the exposed surface: εf=0.7.

� Emissivity of the fire: εf=1.

� Initial temperature: T0=20ºC.

On the other hand, the mechanism that governed the heating through each part of

connection was conductivity, which depends on the thermal properties of the material. In the

case of interactions between surfaces, heat transfer is produced by radiation and mainly

conduction. For the latter mechanism a perfect contact could be assumed, which implies the

same temperature for the points at boundaries. Nevertheless, thermal resistance was likely to

appear and reduce the heat conduction, which is modelled by defining a gap conductance.

Traditionally, authors have neglected this, but the influence on the temperature field is

considerably important in many cases.


14

Gap conductance has been associated with gap clearance [24] or steel tube temperature

[25] in the interface steel concrete of CFST columns. In the reported connections, not only the

interface between steel tube and concrete core exists, but also several more contacts occurred

due to the fastener system. A dependent definition of thermal resistance for each interaction is

complex and would involve many factors at the same time. Therefore, based on the work of

Espinos et al. [1], constant values of gap conductance were evaluated. Finally, perfect contact

was deemed in all contacts, except for interactions sleeve with bolt shank and hole surfaces

(plate and tube) and tube with concrete infill (Fig. 11), where a gap thermal conductance of

200 W/m2K was employed. A sensitivity study was conducted to reach these assumptions as

it is described in section 3.2.2.

For the radiative heat transfer mechanism surface emissivity for both steel and concrete

was 0.7, and the configuration factor was 1.

Results from the nonlinear thermal analysis were temperature curves obtained for each

node of the three dimensional model.

3.2. Validation of the numerical model

The experimental data from the laboratory tests was used for the validation of the

numerical calculation by means of comparison between temperatures obtained through the

FEM and thermocouples measurements.

In Fig. 12a and 12b the overall temperature-time response for the experimentally

controlled locations of bolt is presented, specifically for specimens to CFST columns of series

3. For the three types of connections comprising the series a good correlation was observed

between experiments and FEA calculations. The most noticeable differences were detected

initially for location 1, which was on the exposed surface, probably caused by an uneven


15

distribution of heat in the furnace or inappropiate thermocouple installation. The same

conclusions were obtained for the rest of the series.

In the following section the influence of the thermal properties on the response of high

strength steel of bolts and the thermal resistance in the contacts is discussed, analysing the

accuracy with test results.

3.2.1. Thermal properties of high strength steel of bolts.

Steel

Chemical composition and manufacturing process make high strength steel of bolts

behave different to conventional mild steel, however, in the abscense of experimental data,

the same thermal and mechanical properties were assumed for both. Carbon content affects

the thermal properties of steel, a higher quantity of carbon reduces steel heat conductivity.

Several authors have dealt with the properties of mild steel or structural steel at elevated

temperatures, and also with the mechanical properties of high strength steel in fire, but only

Kodur et al. [22] included the temperature dependent definition of the thermal properties for

high strength bolt steel. They carried out an experimental program to get an insight into the

mechanical and thermal performance of A325 (yield strength fy= 630 MPa and ultimate

strength fu= 830 MPa) and A490 (fy= 895 MPa and fu= 1030 MPa) at elevated temperatures.

Conventional A36 steel (fy= 290 MPa and fu= 500 MPa) was also tested for comparison.

Finally, laboratory data was used to formulate equations for specific heat and thermal

conductivity. Samples were heated up to 735ºC, when the phase change occurs in steel. It was

observed that higher level of carbon, as possessed by A429, reduced slightly the conductivity.

Fig. 13 depicts the thermal properties established by Kodur [22] and EC3 [18] provisions.


16

In this research, the influence of modifications reported by Kodur et al. [22] was assessed

through FEM simulations. Specimens of Series 2 were calculated considering both EC3 [18]

and Kodur et al. [22] thermal properties for bolts M16 grade 8.8 (yield strength fy= 640MPa

and ultimate strength fu= 800 MPa). Eventually, the slight differences in properties meant

there was no significant influence on the connection temperatures.

3.2.2. Gap thermal conductance

A gap conductance coefficient was employed at the boundary of elements in contact to

take into account the resistance to heat transfer conduction across the interactions. Model

interactions were grouped in steel-steel and steel-concrete contacts in order to study its

influence on temperature field.

Steel-steel interactions

Steel-steel interactions occurred between steel elements surfaces of the connection:

bolt head to plate, sleeve to plate hole surface, sleeve to tube hole surface, sleeve to bolt

shank, sleeve to bolt head and plate to tube column. First approach adopted null thermal

resistance in all of them, but results overestimated bolt temperatures, Fig. 14a, that led to

consider a gap conductance at some of the boundaries.

Actually, clearances appeared between some of these surfaces. Most perceptible voids

were located between sleeve and hole surfaces, and also between sleeve and bolt shank.

Concrete mixture filled the spaces, as could be observed in test specimens of Fig. 9. So, a gap

conductance was finally adopted in sleeve to hole surfaces and sleeve to shank interactions.

The recommended constant value [1, 24] of 200 W/m2K was evaluated, which gave proper

results.


17

Fig. 14a shows the comparison between numerical values of temperatures considering

perfect contact and thermal resistance, for HB connection to CFST of series 1, and also with

experimental data. The inclusion of the indicated thermal resistance enhanced the FEM

accuracy. Point 1 (HB1), directly exposed to fire, presented the same temperature with and

without the thermal resistance. However, temperature at points 2 and 3 (HB2 and HB3)

decreased in respect of the case of using perfect contact.

Different values of gap conductance were studied: 100 or 10 W/m2 K, but they did not

substantially affect the results.

Concrete-steel interactions

Concrete-steel interactions included the contact between the steel tube column and

concrete core, and the contacts of the embedded part of fastener system with concrete mixture

around it.

An initial approach assumed a value of 200 W/m2 K, following the guidance of [1, 24]

for all the interactions steel to concrete without distinction. The higher dilatation of steel

involved the separation between the steel tube and concrete infill during fire, as a

consequence, a gap thermal conductance appeared. Nevertheless, in the area of embedded

bolt, dilatation of steel was limited by the concrete, so separation between concrete infill and

steel bolt surfaces resulted consistently difficult to materialize. Therefore, a second approach

considered the elimination of the gap at fastener boundaries. Comparison between both

approaches and tests is presented in Fig. 14b, for HB connection of series 1. The effect of

considering perfect contact at boundaries of the blind-bolt with concrete reduced the

temperature at points 2 and 3, and enhanced the correlation with the experiments.


18

4. NUMERICAL MODEL OF THE WHOLE CONNECTION

Real bolted endplate connections between an I-beam and a CFST column were not tested

in this research, but small-scale connections were tested instead. Validation of the FEM

simulation with the test specimens justified the appropriateness of assumptions considered

and their application to develop a numerical model of the whole bolted connection. Therefore,

with the aim of extending the insight into the thermal behaviour of a single blind-bolt in a

connection to HSS and CFST column, an entire representation of a complete joint was

accomplished by means of FEMs, Fig. 15.

The simulated connection was mainly composed of five different parts as it is presented in

Fig. 16:

� a hollow square steel section column with the same size as the specimens of series

2, i.e. 220 mm in side length and 10 mm in thickness,

� a standard I-shape section steel beam named IPE180,

� four Hollo-bolt fasteners M16 grade 8.8, in two rows separated 100 mm in height.

� a rectangular steel plate with dimensions 200x130x15 mm,

� a concrete slab on the beam, representing real conditions of a building frame,

� concrete infill inside the tube column in connections to CFST.

In order to reduce computational cost and due to symmetry of geometry and thermal load,

only half of the connection was modelled. The same thermal material properties and scheme

for the heat transfer as in the small scale examples were applied. Fire development was

assumed in the storey below, so the exposed areas were all the external surfaces of the

elements except for the upper flange of beam, the concrete slab and the part of column over

the concrete slab. The thickness of the concrete slab is not relevant in this study and only its


19

effect as heat sink and storey separator was taken into account. Nonetheless, to prove the lack

of slab thickness influence, the bolt temperature for slabs 20 and 60 mm thick was compared,

detecting negligible differences.

4.1.1. Results and discussion

The overall temperature-time response for the three locations of the blind-bolts was

obtained, but only results at 30 min will be analysed here due to the similar behaviour patterns

detected along fire exposure.

Fig. 17 compares bolt temperature in the two rows from the endplate connection with

values calculated numerically and experimentally in the small-scale specimens, for the same

connection (HB connections of series 2).

Fig. 17a shows results for the I-beam connection to HSS column. Values calculated in

the reduced numerical model fit accurately the temperature of the bottom row bolts, located

closer to the bottom flange of the beam. Conversely, bolt temperatures in the upper row

registered 100ºC less than in the case of the simple FEM model, due to the influence of

concrete slab effect, which was not included in the small-scale models.

Similar patterns of behaviour were detected in the cases of connection to CFST column

Fig. 17b. The simple model, with one blind-bolt, achieved better correlation with the bolts in

the lower row, which were less affected by the concrete slab effect. In addition, the most

embedded section of the bolt (position 3) was less influenced by concrete slab, and

consequently negligible differences appeared between the simple model and the model of the

whole connection.

In conclusion, bolts in the complete endplate connection registered similar temperature

distribution as in the small-scale models. Therefore, these simple models provided a good


20

approach to representing the thermal behaviour of the blind-bolt in the connection. Moreover,

they made a safe prediction in the case of bolts closer to the concrete slab.

5. EVALUATION OF EC3 PART 1.2 ANNEX D APPROACH

Performance of advanced numerical models to calculate the temperature in connections is

not always practical, thus literature review was accomplished to seek other simple methods.

However, no simple analytical expressions were found suitable for connections to HSS and

CFST, which has not helped the spread of the use of these connections.

This review included the study of Eurocode provisions relating to joints at elevated

temperatures. EC4 [21] gives some requirements that specific connections should comply

with to provide adequate fire resistance, but temperature distribution was not dealt with. On

the other hand, EC3 in Clause D3 of Annex D [18] includes a simple method based on the

behaviour of steelworks, which calculates the temperature in connections where beam

supports a concrete slab. Suitability of this method is evaluated in the following section for

HSS and CFST connection, using FEM of the whole endplate connection for comparison.

5.1. Simple method of EC3 Part 1.2 Annex D. Application

The method consists of equations (1) and (2) that determine the temperature θh in a

certain depth of the connection h (Fig. 16) as a proportion of the beam bottom flange

temperature oat midspan θo. The method can be applied to beam to column and beam to beam

connections. The selection of which equation to use, depends on the height of the beam, D.

� �D 400 0,88 1 0,3 /h o h DT Td �ª º¬ ¼ (1)

D > 400


21

h D/2d 0,88h oT T (2)

h > D/2 � �0,88 1 0,2 1 2 /h o h DT T � �ª º¬ ¼ (3)

Equation (1) was employed to estimate the temperature in an exterior point of the

connections presented in the previous section, between a steel I-beam IPE180 and a square

tube column 220x220 mm. Both, HSS and CFST columns were considered.

Before the application of the thermal gradient, it is necessary to determine temperature in

the exposed bottom flange of the beam θo. For this purpose, and basing the whole calculation

on analytical formulation of Eurocodes, equation (3) included in EC3 Clause 4.2.5.1 [18], was

utilised.

,/m

a t sh neta a

A Vk h tc

TU

' '� (3)

This expression calculates a uniform value of temperature for an unprotected steel

section exposed to fire by means of increments of time Δt. Flux of heat hnet is divided by ca ρa

,that are the specific heat and the unit mass of steel, respectively. The ratio Am/V section

factor indicated the relation of exposed area Am of the element per unit of length divided by V

volume of the element per unit of length in steel sections. Finally, ksh is the correction factor

for the shadow effect, the value of which was calculated following the recommendations of

EC3 [18].

Before obtaining temperature distribution along connections, accuracy of beam

temperature extracted with EC3 Clause 4.2.5.1 [18] was verified against values from the FEM

model, comparing temperature-time curves. Due to the similarity of results and to avoid

confusion, these were not included in Fig. 18.


22

5.2. Results and discussion

Temperatures at two depths of the connection, corresponding with the axis of the two bolt

rows presented in Fig. 16 were studied: the head of blind-bolts in the upper row located at

0.75D from the beam bottom flange, and the ones in the lower row, at 0.25D.

Fig. 18a presents equation results for connections to HSS. In comparison with FEM

calculations, equation (1) of Annex D [18] was conservative up to approximately 20 min of

fire exposure for bolts in the top row, and 23 min for bolts in the bottom row.

Moreover, temperature of HSS column was calculated using equation (3), aiming to know

whether that value could be directly used to approach the temperature of bolts. In connections

to HSS, the temperature of the tube column represented a safe estimation for the bolt in the

bottom row, except for the first 10 minutes of exposure. In the case of blind-bolts in the top

row, the use of tube column values implied that bolt temperature was overestimated by 150ºC

after 30 min of exposure. Fig. 18a also shows a good correlation between analytically and

numerically time-temperature curve for the column.

Fig. 18b depicted results for connections to CFST. EC3 [18] predictions were safe up to

30 min for the top row bolts and 34 min for the bottom row bolts, which meant 10 min more

of fire exposure than in the case of the HSS column.

Equation (3) is not applicable for calculating temperature in CFST section and a

discussion about this is developed in the next section. Fig. 18b includes the time-temperature

curve for the steel tube of the CFST only from the FEM model. Temperatures of the tube

column and blind-bolt from simulations showed slight differences between them. Bolts in the

bottom row were 20-30 ºC hotter than in the tube column because the concrete infill influence


23

was less important. Conversely, temperature in upper bolts was 30-50 ºC lower than

temperature column.

In conclusion, equation of Annex D [18] provides a rough approximation of blind-bolt

temperatures in HSS and CFST connections. Although for CFST connections gave safe

values up to 30 min of fire exposure, it is not able its use as a feasible tool to determine the

temperature in the bolts.

6. DISCUSSION ABOUT OTHER DESIGN RECOMMENDATIONS

Application of other alternative methods to calculate the temperature in the exposed

head of blind-bolts in connections to CFST is presented here. Firstly, the previously

introduced equation of EC3 in clause 4.2.5.1 [18] is analysed, together with its modifications

by Ding and Wang [17]. Thereafter, proposals of Espinos et al. [26] and Leskela [27] are

evaluated.

6.1. Simple calculation method of EC3 Part1.2

The use of equation (3), included in section 5.1 of the present report, required the

definition of the factor Am/V. Table 4.2 of EC3 [18] contains some expressions to calculate

this factor, but always for steelworks. Therefore, equation (3) cannot be applied for a

composite steel and concrete section unless the column is deemed either as a hollow section,

disregarding the effect of concrete (termed as HSS), or as solid steel section (termed as

CFST). Table 2 includes the values of Am/V under these two hypotheses.

EC3 [18] simple method was employed by Ding and Wang [17] to calculate the

temperatures in connections to CFST columns. They developed several expressions to obtain

the section ratio Am/V, which depended on the part of the joint and the type of assembly.


24

Table 2 presents the section factors for endplate bolted connections. The section factor 1 only

considered the endplate being heated from one side and used the thickness of the endplate

(t1). The factor 2 assumed that the combined endplate (t1) and tube thickness (t2) was heated

from one side. Finally, the factor 3 considered the heating corner and used the thickness (t1)

and width of the endplate (l1), and the thickness (t2) and width of the tube (l2).

6.2. Equivalent temperature by Espinos, and Leskela

Espinos et al. [26] and Leskela [27] worked on obtaining an equivalent temperature for

steel and concrete in CFST. Their aim was to create a simple method that provided a uniform

value of temperature for each material of the section, so that, it did not represent a problem to

determine design capacity of the composite element.

The formula proposed by Espinos et al. [26] for the equivalent temperature of steel was

applicable for any fire resistance time. Although it was obtained for circular sections, here it

will be assessed for square and rectangular sections. The value of temperature depended on

the ratio Am/V and the time R.

2, 342.1 10.77 0.044 3.922 / 0.025 /a eq m mR R A V R A VT � � � � � (4)

Leskela [27] provided two tables with equations for circular and rectangular CFST.

Those estimated the steel temperature for different time resistance periods as a function of

diameter and side length (b). For a fire exposure of 30 min, the temperature in the square

section is given by the following equation:

,400650 45

280a eqbT �

� � 120 400bd d (5)


25

6.3. Results and discussion

The suitability of these methods to calculate the external temperature of the blind-bolt

was checked by comparison with temperatures measured in laboratory for position 1 of blind-

bolts. Provided that the type of bolt, Hollo-bolt or Extended Hollo-bolt, did not influence that

temperature, only the first type of fastener was considered. The fire exposure time used as a

reference was 30 min and the four different tube section dimensions were taken into account.

Equation (3) was evaluated assuming 5 different section ratios Am/V for the connection

section: as a hollow section (EC3-HSS), as a solid section of steel (EC3-CFST), and the three

proposals from Ding and Wang [17].

The section factor for the hollow section generated the highest temperature, around

150ºC more than in the tests, Fig. 19. Nonetheless, the adjustment was better than the one

assuming a steel solid section, which underestimated the temperature by 300ºC.

For the connections studied by Ding and Wang [17], factor 3 (Table 2) was the most

suitable one. Conversely, for the examples presented in this research, factor 1 (EC3-F1

WANG) achieved the better approximation, whose difference with tests was not higher than

30ºC, Fig. 19. In the case of section 220x220 mm, the same accuracy was obtained with factor

1 and 3 (EC3-F3 WANG). For the rest of sections, factor 3 miscalculated the temperature in

100ºC less. In all cases, factor 2 (EC3-F2 WANG) provided temperatures around 150ºC lower

than the measured ones.

In conclusion, factor 1 estimated values of temperature higher than factors 2 and 3 so

that it was the most suitable ratio for thicker sections, where concrete effect is smaller and

bolts reached higher temperature. But, there were exceptions associated with the width of the

column and of endplate: when the endplate width (l1) was considerably smaller than the tube


26

side length (l2), factor 3 achieved also good adjustment, as it was pointed out for the column

with dimensions 220x220 mm (series 2).

Fig. 19 shows that temperatures calculated using the equations proposed by Espinos et

al. [26] and Leskela [27]. Both presented similar results and overestimated the temperature in

the exposed part of bolts, 100ºC for the square sections 150x150 mm (series 1) and 80ºC for

the rest.

7. SUMMARY AND CONCLUSIONS

This research provides insight into the effect of different aspects that are involved in the

thermal behaviour of blind-bolt connections.

First, the paper deals with the experimental and numerical thermal study of small-scale

connections to HSS and CFST with Hollo-bolt and Extended Hollo-bolt. An experimental

program comprising 12 specimens was carried out, where thermal response was monitored

throughout the fire. Variables considered were: section dimensions of the tube column

(150x150, 220x220, 250x150, 350x150 mm), type of fastener system (Hollo-bolt or Extended

Hollo-bolt), and type of connection (to HSS or CFST). Slight influence of section dimensions

was detected between the four different sizes studied, less than 30ºC differences throughout

the entire period of fire exposure. Conversely, the effect of concrete filling the steel section

tube was noticeable, even in the exposed position of blind-bolts, where 100ºC lower

temperatures were measured in CFST than in HSS connections. Regarding the type of blind-

bolt, the longer shank embedded in concrete of Extended Hollo-bolt presented a negligible

effect on the external and more damaged part of the bolt.


27

Afterwards, FEM of tested specimens were carried out. They simulated with acceptable

accuracy the heat transfer behaviour occurred in the experiments. The main aspects calibrated

in the model were thermal properties of high strength steel bolt and thermal resistance to heat

flow through the contacts. It was concluded that the use of thermal properties of mild steel for

high strength steel bolts represented a good approximation, as well as the recommendation of

employing a gap conductance of 200 W/m2K for the sleeve interaction with the hole surfaces

and the steel tube with the concrete core.

Secondly, the FEA work included a model of whole endplate connections between an

IPE180 beam and a HSS and a CFST column. In addition to the fuller understanding of the

thermal behaviour of the whole connection, the purpose was also to justify the use of small-

scale connections to study the connections behaviour at elevated temperatures. Similar and

safe predictions resulted from the FEM simulations of small-scale specimens.

Finally, the suitability of some non-specific simple analytical methods to obtain the

bolt temperatures in the connections was studied. On the one hand, thermal gradient from

Annex D of EC3 [18] was deemed a poor approximation, safe up to 20 min and 30 min of fire

exposure for connections to HSS and CFST, respectively. On the other hand, the equation of

EC3 Clause 4.2.5.1 [18], function of ratio Am/V, was found not appropriate to be used in

CFST column connections except for the case of using ratios Am/V (factor 1) proposed by

Ding and Wang [17], which gave more realistic approaches. However, their application

should be carefully considered depending on the elements involved in the connections. Lastly,

it was observed that equations of Espinos et al. [26] and Leskela [27], to calculate equivalent

temperature in CFST connections, overestimated the temperature in the connections by

around 100ºC.


28

In conclusion, the present research includes valuable experimental and numerical data

useful to understand the thermal performance of blind-bolt connections and necessary for its

further thermo-mechanical analysis. Nevertheless, additional work is required to extend the

range of parameters and provide simple expressions for bolt temperatures in HSS and CFST

connections.

ACKNOWLEDGEMENTS

The authors would like to express their sincere gratitude to the Spanish “Ministerio de

Ciencia e Innovación” for the help provided through the Project BIA2012-33144 and the

research scholarship BES-2010-035022.

Acknowledge is expressed to Lindapter International® for providing the fastener

systems to carry out the experiments.

REFERENCES

[1] Espinos A, Romero ML, Hospitaler A. Advanced model for predicting the fire

response of concrete filled tubular columns. Journal of Constructional Steel Research

2010. 66(8-9): 1030-1046.

[2] National Institute of Standards and Techology (NIST). Final Report on the collapse of

the World Trade Center towers. USA: National Institute of Standards and Technology.

2005.

[3] National Institute of Standards and Techology (NIST). Final Report on the collapse of

the World Trade Center building 7. USA: National Institute of Standards and

Technology. 2008.


29

[4] Al-Jabri KS, Burgess IW, Lennon T, Plank RJ. Moment-rotation-temperature curves

for semi-rigid joints. Journal of Constructional Steel Research 2005. 61(3): 281-303.

[5] Al-Jabri KS, Seibi A, Karrech A. Modelling of unstiffened flush end-plate bolted

connections in fire. Journal of Constructional Steel Research 2006. 62(1–2): 151-159.

[6] Wang YC, Dai XH, Bailey CG. An experimental study of relative structural fire

behaviour and robustness of different types of steel joint in restrained steel frames.

Journal of Constructional Steel Research 2011. 67(7): 1149-1163.

[7] Dai XH, Wang YC, Bailey CG. Numerical modelling of structural fire behaviour of

restrained steel beam-column assemblies using typical joint types. Engineering

Structures 2010. 32(8): 2337-2351.

[8] Tizani W, Al-Mughairi A, Owen JS, Pitrakkos T. Rotational stiffness of a blind-bolted

connection to concrete-filled tubes using modified Hollo-bolt. Journal of

Constructional Steel Research 2013. 80(0): 317-331.

[9] Lindapter. Type HB - Hollo-Bolt. Cavity fixings 2,product brochure. Lindapter

International, UK. 2012: p 41-3.

[10] Elghazouli AY, Málaga-Chuquitaype C, Castro JM, Orton AH. Experimental

monotonic and cyclic behaviour of blind-bolted angle connections. Engineering

Structures 2009. 31(11): 2540-2553.

[11] Liu Y, Málaga-Chuquitaype C, Elghazouli AY. Response and component

characterisation of semi-rigid connections to tubular columns under axial loads.

Engineering Structures 2012. 41(0): 510-532.

[12] Liu Y, Málaga-Chuquitaype C, Elghazouli AY. Behaviour of beam-to-tubular column

angle connections under shear loads. Engineering Structures 2012. 42(0): 434-456.


30

[13] Wang ZY, Tizani W, Wang QY. Strength and initial stiffness of a blind-bolt

connection based on the T-stub model. Engineering Structures 2010. 32(9): 2505-

2517.

[14] Pitrakkos T, Tizani W. Experimental behaviour of a novel anchored blind-bolt in

tension. Engineering Structures 2013. 49(0): 905-919.

[15] Research Fund for Coal and Steel (RFCS). Design of composite joints for improved

fire robustness, Final Report. Belgium: European Commission 2014.

[16] Ding J, Wang YC. Experimental study of structural fire behaviour of steel beam to

concrete filled tubular column assemblies with different types of joints. Engineering

Structures 2007. 29(12): 3485-3502.

[17] Ding J, Wang YC. Temperatures in unprotected joints between steel beams and

concrete-filled tubular columns in fire. Fire Safety Journal 2009. 44(1): 16-32.

[18] CEN. EN 1993-1-2, Eurocode 3: Design of steel structures. Part 1-2: General rules –

Structural fire design. Brussels, Belgium: Comité Européen de Normalisation. 2005.

[19] ISO 834: Fire resistance tests, elements of building construction. Switzerland:

International Standards Organisation. 1980.

[20] Pawtucket RIH, Karlsson & Sorenson, Inc. ABAQUS. ABAQUS/Standard Version

6.6 User’s Manual: Volumes I-III.

[21] CEN. EN 1994-1-2, Eurocode 4: Design of composite steel and concrete structures.

Part 1-2: General rules - Structural fire design. Brussels, Belgium: Comité Européen

de Normalisation. 2005.


31

[22] Kodur V, Kand S, Khaliq W. Effect of Temperature on Thermal and Mechanical

Properties of Steel Bolts. Journal of Materials in Civil Engineering 2012. 24(6): 765-

774.

[23] CEN. EN 1991-1-2, Eurocode 1: Actions on structures. Part 1-2: General actions -

Actions on structures exposed to fire. Brussels, Belgium: Comité Européen de

Normalisation. 2002.

[24] Ding J, Wang YC. Realistic modelling of thermal and structural behaviour of

unprotected concrete filled tubular columns in fire. Journal of Constructional Steel

Research 2008. 64(10): 1086-1102.

[25] Ghojel J. Experimental and analytical technique for estimating interface thermal

conductance in composite structural elements under simulated fire conditions.

Experimental Thermal and Fluid Science 2004. 28(4): 347-354.

[26] Espinos A, Romero ML, Hospitaler A. Simple calculation model for evaluating the

fire resistance of unreinforced concrete filled tubular columns. Engineering Structures

2012. 42: 231-244.

[27] Leskela MV. Background Document II: Charateristics for a simple calculation

method/New Annex H/EN 1994-1-2 (C). Draft 16.01.2013. 2013.


32

a)

b)

Fig. 1. Blind-bolts: a) Hollo-bolt type, b) Extended Hollo-bolt type.

Standard bolt

Sleeve

Threaded cone

Standard bolt

Sleeve

Threaded cone

Anchor nut


33

Series 1 150x150 t=8 mm




UHB Hollo-bolt

to HSS

HB Hollo-bolt to CFST

EHB Extended Hollo-bolt to CFST

Fig. 2. Test specimens.

a)

b)

Fig. 3. Test specimens before pouring concrete: a) Hollo-bolt, b) Extended Hollo-bolt.


34

Fig. 4. Location of test specimens in furnace.


35

Fig. 5. Thermocouples position in specimens of series 1.

Fig. 6. Time-temperature response measured by thermocouples in CFST connection with HB of series 3.

0

100

200

300

400

500

600

700

800

900

0 10 20 30 40 50

Tª (

ºC)

Time (min)

FURNACE-Series3HB1HB2HB3HB4HB5HB6HB7HB8HB9HB10HB11


36

Fig. 7. Comparison of bolt temperature between different steel sections in test specimens of HB to CFST connections.

Series 1 Series 2

Series 3 Series 4 0

100

200

300

400

500

600

700

800

900

0 10 20 30 40

Tª (º

C)

Time (min)

FURNACE-Series1HB1-Series1HB2-Series1HB3-Series1FURNACE-Series2HB1-Series2HB2-Series2HB3-Series2FURNACE-Series3HB1-Series3HB2-Series3HB3-Series3FURNACE-Series4HB1-Series4HB2-Series4HB3-Series4

1 2

3

1

2

3

1

2 3

1

2

3


37

a)

b)

Fig. 8. Test temperatures in the three type of connections: a) section 150x150 mm (series 1), b) section 250x150 mm (series 3).

Fig. 9. Specimen after fire exposure.

0

100

200

300

400

500

600

700

800

900

0 10 20 30 40

Tª (

ºC)

Time (min)

FURNACE-Series3UHB1UHB2UHB3HB1HB2HB3EHB1EHB2EHB3

UHB HB EHB

UHB HB EHB

1

2

3

1

2 3

1

2

3

1

2 3

1

2 3

1

2

3

0

100

200

300

400

500

600

700

800

900

0 10 20 30 40

Tª (º

C)

Time (min)

FURNACE-Series1UHB1UHB2UHB3HB1HB2HB3EHB1EHB2EHB3


38

a) Hollow steel section

b) Steel plate

c) Blind-bolt

d) Concrete infill

Fig. 10. Parts of the connection (FEM).

Fig. 11. Gap conductance in interactions.

Sleeve Hollo-bolt Extended Hollo-bolt


39

A

a)

b)

Fig. 12. Bolt temperature in connection to CFST (series 3). Test measurements and FEM model results: a) in HB, b) EHB.

0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60

Tª (

ºC)

Time (min)

FURNACE-Series3

EHB1-EXP

EHB2-EXP

EHB3-EXP

EHB1-NUM

EHB2-NUM

EHB3-NUM

0

100

200

300

400

500

600

700

800

900

1000

0 10 20 30 40 50 60

Tª (

ºC)

Time (min)

FURNACE-Series3

HB1-EXP

HB2-EXP

HB3-EXP

HB1-NUM

HB2-NUM

HB3-NUM


40

Fig. 13. High-temperature thermal properties of high strength steel of bolts. Comparison between EC3 and experimental data from Kodur [22]

a)

b)

Fig. 14. Gap conductance influence on HB temperature in CFST (series 1) in interactions: a) sleeve-hole surfaces, b) concrete-fastener system.

0

10

20

30

40

50

60

0 200 400 600 800 1000 1200

Ther

mal

Con

duct

ivity

(W/m

K)

Temperature (ºC)

EC3A36A325A490

HB

0

100

200

300

400

500

600

700

Tem

pera

ture

ºC

EXPFEM-calibratedFEM-perfect contact in all steel contacts

HB1

HB2

HB3

0

100

200

300

400

500

600

700

Tem

pera

ture

ºC

EXP

FEM-calibrated

FEM-gap in all concrete contacts

HB1

HB2

HB3

0

1000

2000

3000

4000

5000

0 200 400 600 800 1000

Spec

ific

Hea

t (J/

kgºK

)

Temperature (ºC)

EC3A36A325A490


41

Fig. 15. FEM of the endplate connection between an I-beam and the hollow section column.

Fig. 16. Geometrical definition of the connection between the I-beam and the hollow section column.


42

a)

b)

Fig. 17. Bolt temperatures comparison between experimental, FEM of the test specimen and FEM of the whole connection at 30 min of fire exposure: a) in a HSS column, b) in a CFST

column.

a)

b)

Fig. 18. Temperature in column and head of blind-bolts. Results from EC3 and FEM models for connections I-beam to: a) HSS column, b) CFST column.

HB2 HB3 HB1

0

100

200

300

400

500

600

700

800

900

0 10 20 30

Tª (

ºC)

Time (min)

FURNACE-Serie2

HSS-EC3

UHB1-EC3-BOTTOMBOLT

UHB1-EC3-TOPBOLT

HSS-NUM

UHB1-NUM-BOTTOMBOLT

UHB1-NUM-TOPBOLT

UHB1 UHB2 UHB3

0

100

200

300

400

500

600

700

800

900

0 10 20 30

Tª (

ºC)

Time (min)

FURNACE-Serie2

HB1-EC3-BOTTOMBOLT

HB1-EC3-TOPBOLT

CFST-NUM

HB1-NUM-BOTTOMBOLT

HB1-NUM-TOPBOLT

0

100

200

300

400

500

600

700

800

900

Tª (

ºC)

EXP

NUM-1BOLT

NUM-CONNECTION-TOPBOLT

NUM-CONNECTION-BOTTOMBOLT0

100

200

300

400

500

600

700

800

900

Tª (

ºC)

EXP

NUM-1BOLT

NUM-CONNECTION-TOP BOLT

NUM-CONNECTION-BOTTOM BOLT


43

Fig. 19. Temperature in an exposed point of the bolt. Comparison between experiments and simple calculation methods.

0

100

200

300

400

500

600

700

800

Tem

pera

ture

(ºC

)

Series

EXP EC3-HSSEC3-CFST EC3-F1 (WANG)EC3-F2 (WANG) EC3-F3 (WANG)LESKELA ESPINOS

1 2 3 4


44

Table 1. Fire tests specimens.

Specimen index Type of bolt Shank Length (mm) Type of section

Series 1-Section 150x150 t=8 mm UHB16-8.8D-150x150x8 Hollo-bolt 75 HSS HB16-8.8D-C30-150x150x8 Hollo-bolt 75 CFST EHB16-8.8D-C30-150x150x8 Extended Hollo-bolt 120 CFST Series 2-Section 220x220 t=10 mm UHB16-8.8D-220x220x10 Hollo-bolt 75 HSS HB16-8.8D-C30-220x220x10 Hollo-bolt 75 CFST EHB16-8.8D-C30-220x220x10 Extended Hollo-bolt 120 CFST Series 3-Section 250x150 t=10 mm UHB16-8.8D-250x150x10 Hollo-bolt 75 HSS HB16-8.8D-C30-250x150x10 Hollo-bolt 75 CFST EHB16-8.8D-C30-250x150x10 Extended Hollo-bolt 120 CFST Series 4-Section 350x150 t=10 mm UHB16-8.8D-350x150x10 Hollo-bolt 75 HSS HB16-8.8D-C30-350x150x10 Hollo-bolt 75 CFST EHB16-8.8D-C30-350x150x10 Extended Hollo-bolt 120 CFST

Table 2. Section Factors.

SECTION FACTOR 150x150x8 220x220x10 250x150x10 350x150x10

HSS A/VHSS 131.02 103.85 104.27 103.35

CFST A/VCFST 26.45 18.04 21.12 18.91

Factor 1 (Ding and Wang) 1/t1 66.67 66.67 66.67 66.67

Factor 2 (Ding and Wang) 1/(t1+t2) 43.48 40.00 40.00 40.00

Factor 3 (Ding and Wang) l2/(t1l1+t2l2) 52.63 57.14 47.62 47.62


45

LIST OF FIGURE CAPTIONS

Fig. 1 Blind-bolts: a) Hollo-bolt type, b) Extended Hollo-bolt type

Fig. 2 Test specimens

Fig. 3 Test specimens before pouring concrete: a) Hollo-bolt, b) Extended Hollo-bolt

Fig. 4 Location of test specimens in furnace

Fig. 5 Thermocouples position in specimens of series 1

Fig. 6 Time-temperature response measured by thermocouples in CFST connection with

HB of series 3

Fig. 7 Comparison of bolt temperature between different steel sections in test specimens

of HB to CFST connections

Fig. 8 Test temperatures in the three type of connections: a) section 150x150 mm (series

1), b) section 250x150 mm (series 3)

Fig. 9 Specimen after fire exposure.

Fig. 10 Parts of the connection (FEM).

Fig. 11 Gap conductance in interactions.

Fig. 12 Bolt temperature in connection to CFST (series 3). Test measurements and FEM

model results: a) in HB, b) EHB.

Fig. 13 High-temperature thermal properties of high strength steel of bolts. Comparison

between EC3 and experimental data from Kodur [22]

Fig. 14 Gap conductance influence on HB temperature in CFST (series 1) in interactions:

a) sleeve-hole surfaces, b) concrete-fastener system

Fig. 15 FEM of the endplate connection between an I-beam and the hollow section column.


46

Fig. 16 Geometrical definition of the connection between the I-beam and the hollow

section column.

Fig. 17 Bolt temperatures comparison between experimental, FEM of the test specimen and

FEM of the whole connection at 30 min of fire exposure: a) in a HSS column, b) in a CFST

column.

Fig. 18 Temperature in column and head of blind-bolts. Results from EC3 and FEM

models for connections I-beam to: a) HSS column, b) CFST column.

Fig. 19 Temperature in an exposed point of the bolt. Comparison between experiments and

simple calculation methods.

LIST OF TABLES

Table 1 Fire tests specimens

Table 2 Section Factors.

View publication statsView publication stats

https://www.researchgate.net/publication/272022252

Pascual, Ana M. and Romero, Manuel and Tizani, Walid (2015...

Documents

Transcript of Pascual, Ana M. and Romero, Manuel and Tizani, Walid (2015...